marinaabakarova/proteocast repository overview

⁠ProteoCast Docker image

This repository provides a Docker image for the ProteoCast pipeline⁠.

⁠Docker

⁠First, you need to install Docker from Docker website⁠.

⁠Pull the image

⁠To download ProteoCast image use the following command line:

docker pull marinaabakarova/proteocast:1.0

⁠Run ProteoCast locally

ProteoCast is distributed as a Docker image and can be run locally on any system with Docker installed (macOS, Linux, Windows).

ProteoCast can be run in two modes:

1. File mode: you provide the MSA file (and optionally a PDB file)
2. UniProt mode: you provide a UniProtID and the pipeline retrieves the required resources from AlphaFoldDB when applicable.

⁠1. Prepare a job directory

Create a directory that will be mounted inside the container. If you prefer to be in file mode, this directory must contain:

• a multiple sequence alignment (MSA) file (.a3m or .fasta)
• and optionally a PDB structure

⁠2. Create a run script

Inside the job directory, create a bash script (e.g. nano proteocast.sh):

#!/usr/bin/env bash
set -euo pipefail

JOB_DIR="/ABSOLUTE/PATH/TO/JOB_DIRECTORY"

# Arguments:
#   $1 : input MSA file (.a3m or .fasta) or UniProt code
#   $2 : chain identifier
#   $3 : UniProt ID (optional; used to retrieve AlphaFoldDB structure if no PDB is provided)
#		if using a custom pdb structure just put it in the working folder, no argument needed
docker run --rm \
  -v "${JOB_DIR}:/opt/job" \
  marinaabakarova/proteocast:latest \
  /bin/bash -c "cd / && bash run.sh $1 $2 ${3:-}" \
  2>> "${JOB_DIR}/proteocast.log"

⁠3. Make the script executable:

chmod +x proteocast.sh

⁠4. Run ProteoCast

From inside the job directory, run

./proteocast.sh <msa_file> [chain][uniprot_id]

⁠Examples

Run with an uploaded MSA and a local PDB file

./proteocast.sh test_3782f.a3m A

Run with an uploaded MSA and retrieve the structure from AlphaFoldDB

./proteocast.sh test_3782f.a3m A Q9UET6

Run using a UniProt ID with automatic retrieval of AlphaFoldDB files

./proteocast.sh P01308 A P01308

Standard error logs are written to: proteocast.log

⁠Feedback/Issues

Please report any issues or feedback to [email protected]⁠